High Performance Reversible Data Hiding for Mobile Applications and Human Interaction

Reversible data hiding with predictive error-based difference expansion to conceal personal and sensitive information in mobile applications and human interaction activities through hand-held devices recently has drawn lots of interest. It can not only completely restore the original image after extraction, but also keep the protected privacy information imperceptible against the eavesdroppers. A reversible data hiding method based on prediction-error and pixel-value difference was proposed in this paper. The suitable situations for reversible algorithms and embedded value are selected depended on the differences between image regions and the characteristics of embedded data. Experimental results show that our proposed scheme provides a 57956 bits (0.24 ppb) hiding capacity and a 53.81dB PSNR on F16 image without generating noticeable visual artifacts. The proposed method is additionally applied with different predictors.